v0.9.5

• Fix a bug in which default cluster thresholds are calculated incorrectly when
  the optimized thresholds are not given as fractional distances (as opposed to,
  e.g., percent identity).
• Fix use of id_col to specify a non-default column name for sequence IDs in
  the taxonomy argument to optimize_thresholds().